Can prior authorization be fixed? Should it be?

Controlling health care costs depends on eliminating unnecessary, low-value care. But insurers have an inherent conflict of interest and shouldn't be running prior authorization.

Anger over prior authorization, where private insurers review and sometimes deny health care service requests by hospitals and physicians, has reached the boiling point.

Providers oppose its use, citing the huge documentation burden and frequent delays in service delivery. Patients are also feeling the pain, raising it to their second biggest problem (after affordability) when accessing care.

A KFF poll released last year found over half of insured adults say their physicians’ proposed care plan has been subject to prior authorization in the past two years. Over half of them find it difficult to navigate the process.

Insurers’ increased use of artificial intelligence (AI) when making prior authorization reimbursement decisions is also drawing fire. It began when a Stat News investigation in 2023 concluded UnitedHealth Group ordered employees to rigorously follow an AI algorithm when evaluating the need for post-acute care for its Medicare Advantage members, which sharply reduced utilization.

A subsequent Senate investigation confirmed that finding. “United Healthcare’s prior authorization denial rate for post-acute care surged from 10.9 percent in 2020 to 22.7 percent in 2022,” the report noted. “During this time, it was implementing multiple initiatives to automate the process.”

Despite those concerns, prior authorization requests are rising almost as fast as the server farms fueling the computers that are making the decisions. A recent KFF issue brief found Medicare Advantage insurers, who now cover more than half of seniors, denied 7.4% of 53 million service requests they reviewed in 2024, a sharp increase from the 6.4% of 50 million requests denied the previous year.

Legislators across the country are responding to the backlash. Last December, Rep. Greg Landsman (D-OH) and Rep. Bonnie Watson Coleman (D-NJ) introduced legislation that would ban the use of AI in making coverage determinations in Medicare. Stateline has reported at least four states — Arizona, Maryland, Nebraska and Texas — now prohibit insurance companies from using AI as the sole source for making a service payment denial.

The Trump administration’s senior health care leadership, meanwhile, has taken its usual voluntary approach to seeking change from the insurance industry, despite having made fighting fraud a centerpiece of its attacks on Medicaid and the Affordable Care Act’s exchange-based plans. CMS administrator Mehmet Oz frequently touted the superiority of Medicare Advantage plans when a television celebrity, and recently said he’d like to see more beneficiaries join those plans.

In a press conference held last June featuring the president, HHS Secretary Robert F. Kennedy Jr., and Oz, the administration announced it had reached a voluntary agreement with a half dozen insurers and industry trade associations, who promised to streamline the prior authorization process and limit the number of services subject to it. They said the insurers’ plans would be announced by January 1, 2026.

This week, I sent emails to a half dozen major insurers asking about their plans. None have responded substantively. A spokesman for AHIP, the industry trade group, said “work is underway to create greater standardization” and “we will be sharing updates this spring.”

Is this really WISeR?

CMS’ leaders can’t be too upset with prior authorization. The Innovation Center’s new leader, Abe Sutton, has moved forward with a prior authorization pilot project that will subject some seniors in traditional Medicare to the process. The project, dubbed WISeR (Wasteful and Inappropriate Service Reduction), allows the use of AI in making coverage determinations.

But in a nod to its critics, the pilot project, which will affect traditional Medicare beneficiaries in Arizona, New Jersey, Ohio, Oklahoma, Texas, and Washington, will limit its use to 17 high-volume services prone to abuse. They include some back surgeries, knee arthroscopy and skin substitutes. The pilot project will also require every computer-proposed denial be reviewed by a clinician.

There’s no doubt that unnecessary utilization and wasteful spending are a serious problem in the U.S. health care system, even as it shortchanges many necessary and high-value preventive measures. Numerous studies have estimated anywhere from 20% to 30% of all health care spending could be eliminated without detracting from health.

But bulk of waste doesn’t come from overtreatment. It includes wildly overpriced drugs, tests and procedures (when compared to their actual medical value); exorbitant administrative spending by insurers and providers to maintain their complex billing systems; a lack of care coordination; and outright fraud. None of those issues are addressed by prior authorization. Those are best managed by, respectively, more hard-nosed price negotiations by payers; payment reform; increased spending on primary care; and more rigorous fraud enforcement by the Department of Justice, the HHS Office of the Inspector General, and states’ Attorneys General.

The unnecessary care that prior authorization is meant to address accounts for just 6% to 8% of spending. While that seems like a low percentage, in a $5 trillion-a-year system, that’s about $350 billion a year in potential savings that could used to either immediately lower costs or spend on the prevention measures and primary care that lower costs in the long run.

Where can that overtreatment be found? It often involves highly-paid physicians padding their incomes by conducting tests, procedures and operations that cannot be justified by their specialties’ clinical practice guidelines. Or, it comes in the gray areas of clinical decision-making where there are lower-cost approaches that are equally effective, or there is good reason to think a particular patient might benefit even though they don’t technically meet the guidelines’ criteria.

These real world considerations demand that any utilization management tool aimed at rooting out unnecessary care be wielded with scalpel-like precision, and not be driven by algorithm-driven computers. It also must be put in the hands of people who do not profit from the final determinations. The private insurers who use prior authorization earn more money when they deny care.

Does the new WISeR model for traditional Medicare eliminate that conflict of interest?

To its credit, WISeR prohibits AI-generated reviews from being the last word in a service denial. The computer-generated review can be used to approve a service. But any proposed denial must be reviewed by a clinician.

However, WISeR failed to eliminate financial conflicts of interest from the process. It contracted with six outside vendors to conduct the reviews, one for each state. They were chosen because they are “technology companies with experience managing prior authorization programs for other payers, including Medicare Advantage (MA),” according to Applied Policy, an industry consulting firm.

They will earn their keep in the same way that they make money for their MA clients. Their reimbursement from CMS will come in the form of a contingency fee based on “the Medicare savings generated when designated services are not affirmed and ultimately not paid.”

In other words, the more they deny, the more they earn. “Such a structure creates a perverse incentive to deny care that otherwise may be appropriate, as vendors may increase their profits by denying care,” the American Hospital Association said in a letter to CMS.

Published reports say the prior authorization review vendors will earn somewhere between 10% and 20% of the cost of every service they deny. “Adding prior authorization requirements to traditional Medicare will create costly problems and barriers to necessary care,” the Center for Medicare Advocacy said in a statement shortly after the program was announced. “Instead, to truly address waste and abuse in Medicare, CMS should look to the dramatic overpayments and unreasonable denials in Medicare Advantage.”

To fix WISeR, CMS should revamp its reimbursement model so that vendors receive a small flat fee for every prior authorization request evaluated by their computers and a larger flat fee for every computer-generated denial adjudicated by its salaried clinicians. Payer savings (in this case Medicare) should never enter into the equation.

There is a better way

But even a WISeR free of conflicts of interest begs the larger question of whether prior authorization is the most effective tool for eliminating unnecessary care. In my view, putting providers on annual budgets and letting them make those decisions in-house is the better way to go.

It eliminates the middle-man in clinical decision-making. It encourages physicians, who are increasingly salaried and employed by large organizations, to adhere more closely to accepted clinical practice guidelines. It enhances professional autonomy, which should reduce burnout and improve the physician-patient relationship.

Prior authorization proponents argue that putting providers on budgets will lead to rationing. Organizations will simply stop offering high-priced care as they near the end of the year with exhausted budgets.

That presumes hospitals and physician practices are incapable of changing their practice patterns; making more rational treatment decisions; devoting more resources to primary care; and, through those actions, reduce overall spending. That’s ultimately where health care needs to go if it is going to prevent the expensive hospitalizations that drive total costs.

Fixed budgets enables provider organizations to flexibly deploy their resources in ways that achieve those goals. Prior authorization, even if streamlined, well-targeted and free of conflicts of interest, may reduce costs, but it offers no guarantee that it will lead to better health.